Remote Cluster Status View shows status as "initing" and the Heartbeat Status is "Icmp: ALIVE" after connecting two or more clusters into a global cluster

Remote Cluster Status View shows status as "initing" and the Heartbeat Status is "Icmp: ALIVE" after connecting two or more clusters into a global cluster

  • Article ID:100018333
  • Last Published:
  • Product(s):InfoScale & Storage Foundation

Problem

After the successful completion of the Add/Delete Remote Cluster Wizard to configure the Storage Foundation for Windows (tm) High Availability (SFW-HA) Global Cluster Option (GCO) to connect two or more clusters,  the Remote Cluster Status View shows different values in the different clusters.

The Remote Cluster Status View at the primary site shows the Status of the remote cluster is Initing and the Heartbeat Status is "Icmp: ALIVE" (figure 1).


Figure 1
 

The Remote Cluster Status View at the remote site shows the Status of the remote cluster is Initing and the Heartbeat Status is "Icmp: UNKNOWN" (figure2).

Figure 2
 

Note: The Cluster Names and IP addresses listed above are examples only and should be replaced to match the configuration of the environment.

The GCO cluster heartbeat relies on the ping command to send Internet Control Message Protocol (ICMP) "echo requests" to the remote cluster and listens for ICMP "echo response" replies to validate that the clusters can communicate. The above issue occurs because the network at the primary site has a firewall enabled that is blocking ICMP requests.

Error Message

 

The following errors are shown in the wac_A.txt log, assuming that there is a firewall restriction for ICMP packets to the remote DR site, the PROD cluster heartbeats are down, and the WAC on the PROD site will refuse heartbeat packes from the DR site. 

PROD wac_A.txt

VCS NOTICE V-16-3-18321 All heartbeats from cluster (<prod>) to cluster (<dr>) are down. will not accept connection from cluster (<dr>) until at least one heartbeat to cluster (<dr>) is alive

DR wac_A.txt 

VCS WARNING V-16-3-18325 Cluster (<prod>) refused connection. This is because all heartbeats from cluster (<prod>) to cluster (<dr>) are currently down.

where:

  • PROD refers to the production or primary site and cluster
  • DR refers to the disaster recovery site and cluster
  • WAC is the Wide Area Connector for the GCO

Cause

Firewall is blocking ICMP, port 14155

Solution

To resolve the issue, the IP address of the cluster must be configured to send and receive ICMP packets so it can communicate with the cluster at the remote site. This may involve opening ports on the firewall or changing the IP address of the cluster to a network that already allows ICMP packets.
 

Applies To

GCO cluster with firewall affecting ICMP traffic

References

UMI : V-16-3-18321
UMI : V-16-3-18325

Was this content helpful?